AI Platform

AI Inference served in the EU with zero data retention.

Hosted models behind one OpenAI-compatible API. Keys, usage and privacy controls live in the same project as your app.

# any OpenAI client works, only the base URL changes
curl https://api.nodion.ai/v1/chat/completions \
  -H "Authorization: Bearer $NODION_AI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "aleph-alpha/kolibri-1",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.nodion.ai/v1",
    api_key=os.environ["NODION_AI_KEY"],
)
reply = client.chat.completions.create(
    model="aleph-alpha/kolibri-1",
    messages=[{"role": "user", "content": "Hello"}],
)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.nodion.ai/v1",
  apiKey: process.env.NODION_AI_KEY,
});
const reply = await client.chat.completions.create({
  model: "aleph-alpha/kolibri-1",
  messages: [{ role: "user", content: "Hello" }],
});

Models

The catalog.

Text, images, audio and documents behind one API. Pay only for what you call.

Prices in EUR excl. VAT

Preview models can leave the catalog at short notice. There is no deprecation date.

Text and reasoning

Kolibri 1Previewaleph-alpha/kolibri-1

Input0,15 €Output0,60 €

This model is served through the following providerLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
Nodion	Input0,15 €	Output0,60 €	Latency	Throughput	Uptime

DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flash

Input0,20 €0,01 € cachedOutput1,00 €

This model is served through the following 4 providersLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
TensorX	Input0,50 €0,13 € cached	Output1,50 €	Latency	Throughput	Uptime
Melious AI	Input0,20 €0,01 € cached	Output1,00 €	Latency	Throughput	Uptime
Arkane	Input0,20 €0,01 € cached	Output1,00 €	Latency	Throughput	Uptime
AKI.IO	Input0,25 €0,10 € cached	Output1,00 €	Latency	Throughput	Uptime

Gemma 4 31Bgoogle/gemma4-31b

Input0,25 €0,15 € cachedOutput0,50 €

This model is served through the following 3 providersLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
Melious AI	Input0,10 €0,02 € cached	Output0,30 €	Latency	Throughput	Uptime
Berget AI	Input0,25 €0,15 € cached	Output0,50 €	Latency	Throughput	Uptime
Arkane	Input0,25 €0,15 € cached	Output0,50 €	Latency	Throughput	Uptime

Mistral Large 4Previewmistral/mistral-large-4

Input0,50 €0,07 € cachedOutput1,75 €

This model is served through the following providerLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
Mistral AI (EU)	Input0,50 €0,07 € cached	Output1,75 €	Latency	Throughput	Uptime

Mistral Small 4mistral/mistral-small-4

Input0,15 €0,15 € cachedOutput0,55 €

This model is served through the following 2 providersLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
Mistral AI (EU)	Input0,15 €0,15 € cached	Output0,55 €	Latency	Throughput	Uptime
AKI.IO	Input0,20 €0,15 € cached	Output0,60 €	Latency	Throughput	Uptime

Kimi K3moonshot/kimi-k3

Input3,00 €0,30 € cachedOutput15,00 €

This model is served through the following 3 providersLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
Berget AI	Input3,00 €0,30 € cached	Output15,00 €	Latency	Throughput	Uptime
Arkane	Input3,00 €0,30 € cached	Output15,00 €	Latency	Throughput	Uptime
TensorX	Input3,00 €0,30 € cached	Output15,00 €	Latency	Throughput	Uptime

Qwen3.8 27Bqwen/qwen3.8-27b

Input0,40 €0,03 € cachedOutput2,40 €

This model is served through the following 5 providersLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
Berget AI	Input0,40 €0,03 € cached	Output2,40 €	Latency	Throughput	Uptime
TensorX	Input0,40 €0,03 € cached	Output2,40 €	Latency	Throughput	Uptime
Melious AI	Input0,40 €0,03 € cached	Output2,40 €	Latency	Throughput	Uptime
Arkane	Input0,40 €0,03 € cached	Output2,40 €	Latency	Throughput	Uptime
AKI.IO	Input0,30 €0,10 € cached	Output2,20 €	Latency	Throughput	Uptime

GLM 5.3z-ai/glm-5.3

Input1,00 €0,20 € cachedOutput3,00 €

This model is served through the following 5 providersLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
TensorX	Input1,00 €0,20 € cached	Output3,00 €	Latency	Throughput	Uptime
Melious AI	Input1,00 €0,20 € cached	Output3,00 €	Latency	Throughput	Uptime
Arkane	Input1,00 €0,20 € cached	Output3,00 €	Latency	Throughput	Uptime
Mistral AI (EU)	Input1,00 €0,20 € cached	Output3,00 €	Latency	Throughput	Uptime
AKI.IO	Input1,00 €0,25 € cached	Output3,50 €	Latency	Throughput	Uptime

GLM 5.3 Flashz-ai/glm-5.3-flash

Input0,25 €0,03 € cachedOutput0,50 €

This model is served through the following 3 providersLatency, throughput and uptime over the last 24 hours

Provider	Input	Output	Latency	Throughput	Uptime
Berget AI	Input0,25 €0,03 € cached	Output0,50 €	Latency	Throughput	Uptime
TensorX	Input0,25 €0,03 € cached	Output0,50 €	Latency	Throughput	Uptime
Melious AI	Input0,10 €0,02 € cached	Output0,40 €	Latency	Throughput	Uptime

per 1M tokens

Speech

Voxtral Mini TTSmistral/voxtral-mini-tts

Price13,95 €-

per 1M characters

Document OCR

FPDF Inspectorfirecrawl/pdf-inspector

Max pages500Price0,20 €

Mistral OCR 4.1mistral/mistral-ocr-4-1

Max pages100Price3,50 €

per 1,000 pages

Tools

Web Fetchweb-fetch

Price1,00 €

Web Newsweb-news

Price2,00 €

Web Searchweb-search

Price2,00 €

per 1,000 calls

Prices from the live catalog. Usage and request logs are in the dashboard.

Built for production

Inference you can put in front of customers.

Drop-in

OpenAI-compatible.

Point any OpenAI SDK at api.nodion.ai/v1. Chat, tools, JSON mode and streaming work as you expect.

Keys

Limits per key.

Optional expiry and a daily spend limit on every key. Only project admins create or revoke them.

Privacy

Your data, your rules.

Restrict which regions may serve a project, and turn on zero data retention in the project settings.

Billing

Prepaid wallet.

AI spend comes from a wallet you top up by card or PayPal, separate from the platform invoice.

Give your agent a cloud.

Earn up to 15 € in credit when signing up.

Continue with GitHub

or with email

AI Inference served in the EU with zero data retention.

The catalog.

Text and reasoning

Image generation

Transcription

Speech

Embeddings

Rerank

Document OCR

Tools

Inference you can put in front of customers.

OpenAI-compatible.

Limits per key.

Your data, your rules.

Prepaid wallet.

Give your agent a cloud.